Nuria Oliver

Image analogies

This paper describes a new framework for processing images by example, called “image analogies.” The framework involves two stages: a design phase, in which a pair of images, with one image purported to be a “filtered” version of the other, is presented as “training data”; and an application phase, in which the learned filter is applied to some new target image in order to create an “analogous” filtered result. Image analogies are based on a simple multi-scale autoregression, inspired primarily by recent results in texture synthesis. By choosing different types of source image pairs as input, the framework supports a wide variety of “image filter” effects, including traditional image filters, such as blurring or embossing; improved texture synthesis, in which some textures are synthesized with higher quality than by previous approaches; super-resolution, in which a higher-resolution image is inferred from a low-resolution source; texture transfer, in which images are “texturized” with some arbitrary source texture; artistic filters, in which various drawing and painting styles are synthesized based on scanned real-world examples; and texture-by-numbers, in which realistic scenes, composed of a variety of textures, are created using a simple painting interface.

Coupled hidden Markov models for complex action recognition

We present algorithms for coupling and training hidden Markov models (HMMs) to model interacting processes, and demonstrate their superiority to conventional HMMs in a vision task classifying two-handed actions. HMMs are perhaps the most successful framework in perceptual computing for modeling and classifying dynamic behaviors, popular because they offer dynamic time warping, a training algorithm and a clear Bayesian semantics. However the Markovian framework makes strong restrictive assumptions about the system generating the signal-that it is a single process having a small number of states and an extremely limited state memory. The single-process model is often inappropriate for vision (and speech) applications, resulting in low ceilings on model performance. Coupled HMMs provide an efficient way to resolve many of these problems, and offer superior training speeds, model likelihoods, and robustness to initial conditions.

Multiverse recommendation: n-dimensional tensor factorization for context-aware collaborative filtering

Context has been recognized as an important factor to consider in personalized Recommender Systems. However, most model-based Collaborative Filtering approaches such as Matrix Factorization do not provide a straightforward way of integrating context information into the model. In this work, we introduce a Collaborative Filtering method based on Tensor Factorization, a generalization of Matrix Factorization that allows for a flexible and generic integration of contextual information by modeling the data as a User-Item-Context N-dimensional tensor instead of the traditional 2D User-Item matrix. In the proposed model, called Multiverse Recommendation, different types of context are considered as additional dimensions in the representation of the data as a tensor. The factorization of this tensor leads to a compact model of the data which can be used to provide context-aware recommendations. We provide an algorithm to address the N-dimensional factorization, and show that the Multiverse Recommendation improves upon non-contextual Matrix Factorization up to 30% in terms of the Mean Absolute Error (MAE). We also compare to two state-of-the-art context-aware methods and show that Tensor Factorization consistently outperforms them both in semi-synthetic and real-world data - improvements range from 2.5% to more than 12% depending on the data. Noticeably, our approach outperforms other methods by a wider margin whenever more contextual information is available.

Layered representations for learning and inferring office activity from multiple sensory channels

We present the use of layered probabilistic representations for modeling human activities, and describe how we use the representation to do sensing, learning, and inference at multiple levels of temporal granularity and abstraction and from heterogeneous data sources. The approach centers on the use of a cascade of Hidden Markov Models named Layered Hidden Markov Models (LHMMs) to diagnose states of a users activity based on real-time streams of evidence from video, audio, and computer (keyboard and mouse) interactions. We couple these LHMMs with an expected utility analysis that considers the cost of misclassification. We describe the representation, present an implementation, and report on experiments with our layered architecture in a real-time office-awareness setting.

Layered representations for human activity recognition

We present the use of layered probabilistic representations using hidden Markov models for performing sensing, learning, and inference at multiple levels of temporal granularity We describe the use of representation in a system that diagnoses states of a users activity based on real-time streams of evidence from video, acoustic, and computer interactions. We review the representation, present an implementation, and report on experiments with the layered representation in an office-awareness application.

Graphical models for driver behavior recognition in a SmartCar

In this paper we describe our SmartCar testbed: a real-time data acquisition system and a machine learning framework for modeling and recognizing driver maneuvers at a tactical level, with special emphasis on how the context affects the drivers performance. The perceptual input is multimodal: four video signals capture the contextual traffic, the drivers head and the drivers viewpoint; and a real-time data acquisition system records the cars brake, gear, steering wheel angle, speed and acceleration throttle signals. Over 70 drivers have driven the SmartCar for 1.25 hours in the greater Boston area. Graphical models, HMM and coupled HMM, have been trained using the experiment driving data to create models of seven different driver maneuvers: passing, changing lanes right and left, turning right and left, starting and stopping. We show that, on average, the predictive power of our models is of 1 second before the maneuver starts taking place. Therefore, these models would be essential to facilitate operating mode transitions between driver and driver assistance systems, to prevent potential dangerous situations and to create more realistic automated cars in car simulators.

HealthGear: a real-time wearable system for monitoring and analyzing physiological signals

We present HealthGear, a real-time wearable system for monitoring, visualizing and analyzing physiological signals. HealthGear consists of a set of noninvasive physiological sensors wirelessly connected via Bluetooth to a cell phone which stores, transmits and analyzes the physiological data, and presents it to the user in an intelligible way. In this paper, we focus on an implementation of HealthGear using a blood oximeter to monitor the users blood oxygen level and pulse while sleeping. We also describe two different algorithms for automatically detecting sleep apnea events, and illustrate the performance of the overall system in a sleep study with 20 volunteers

LAFTER: lips and face real time tracker

This paper describes an active-camera real-time system for tracking, shape description, and classification of the human face and mouth using only an SGI Indy computer. The system is based on use of 2-D blob features, which are spatially-compact clusters of pixels that are similar in terms of low-level image properties. Patterns of behavior (e.g., facial expressions and head movements) can be classified in real-time using Hidden Markov Model (HMM) methods. The system has been tested on hundreds of users and has demonstrated extremely reliable and accurate performance. Typical classification accuracies are near 100%.

CLiMF: learning to maximize reciprocal rank with collaborative less-is-more filtering

In this paper we tackle the problem of recommendation in the scenarios with binary relevance data, when only a few (k) items are recommended to individual users. Past work on Collaborative Filtering (CF) has either not addressed the ranking problem for binary relevance datasets, or not specifically focused on improving top-k recommendations. To solve the problem we propose a new CF approach, Collaborative Less-is-More Filtering (CLiMF). In CLiMF the model parameters are learned by directly maximizing the Mean Reciprocal Rank (MRR), which is a well-known information retrieval metric for measuring the performance of top-k recommendations. We achieve linear computational complexity by introducing a lower bound of the smoothed reciprocal rank metric. Experiments on two social network datasets demonstrate the effectiveness and the scalability of CLiMF, and show that CLiMF significantly outperforms a naive baseline and two state-of-the-art CF methods.

I Like It... I Like It Not: Evaluating User Ratings Noise in Recommender Systems

Recent growing interest in predicting and influencing consumer behavior has generated a parallel increase in research efforts on Recommender Systems. Many of the state-of-the-art Recommender Systems algorithms rely on obtaining user ratings in order to later predict unknown ratings. An underlying assumption in this approach is that the user ratings can be treated as ground truth of the users taste. However, users are inconsistent in giving their feedback, thus introducing an unknown amount of noise that challenges the validity of this assumption. In this paper, we tackle the problem of analyzing and characterizing the noise in user feedback through ratings of movies. We present a user study aimed at quantifying the noise in user ratings that is due to inconsistencies. We measure RMSE values that range from 0.557 to 0.8156. We also analyze how factors such as item sorting and time of rating affect this noise.